1. Field of the Invention
The present invention generally relates to an operation processing system, and more particularly, to a server auxiliary operating system able to increase the computing capability of the server.
2. Description of Related Art
Along with the increasing programmable capability and performance of a graphic processing unit (GPU), the application developers have been hoping that the graphic hardware is able to solve the high-density computation jobs which are accomplished by a central processing unit (CPU) before. The GPU executes codes in parallel mode of single instruction multiple data (SIMD), and the code for vector addition extracts two elements to be added from a memory and then performs the vector addition according to the positions of fragments. Meanwhile, the code allocates the result with the output colour. An output memory stores the vector addition, so that the value of the vector addition can be arbitrarily used in the next computation. Thanks for the super-intensive huge parallel computation function of a GPU, the computation capability thereof is far beyond the CPU and other processors. As a result, the GPU serves not only as a plotting engine, but also as the major computing engine of a PC in future.
The GPU card usually serves as an external adaptor card of a server and is disposed in the server case. However, limited by the server case space and the interface number, usually only a certain number of the GPU cards are supported, which limits expanding the server performance. How to implement an appropriate match between a server and GPU cards and expand the quantity of the GPU cards according to the demand of the clients so as realize a flexible configuration have become a significant issue to be solved by the manufacturers.